A randomized, double-blinded controlled trial had been conducted in two hospitals and included patients undergoing valve surgery with or without coronary artery bypass graft (CABG). Patients were randomized to receive (when needed) standard buff y-coat depleted or pre-storage leukoreduced red blood cells. Patient samples were collected aft er informed
consent was obtained. Th e endpoints of the study were: postoperative infections, MODS
and mortality (90-days and in-hospital). Th e design and clinical outcome of the study has
been described elsewhere [17]. In this trial infections were defi ned according to the criteria
of the Center for Disease Control (CDC) [20]. Th e following infections were taken into
account: respiratory tract, urinary tract, sepsis and wound infections. Th e incidence and
duration of organ dysfunction was described as defi ned by Knaus [21]. MODS was defi ned
as failure of two or more organ systems. Th e blood products used in this study were prepared
and controlled according to the Dutch standards for blood banks.
Blood samples were taken pre-operatively before the start of the surgical procedure and at admission on the ICU, immediately centrifuged, and the sera stored at -80oC.
Measurement of MBL
Th e concentration of MBL was measured by sandwich enzyme linked immunosorbent assay
(ELISA) as described previously [22]. In summary, 96-well ELISA plates were coated with
mAb 3E7 (monoclonal anti-MBL antibody kindly provided by Dr. T. Fujita, Fukushima, Japan). Aft er blocking residual binding sites with PBS containing 1% BSA and washing, serum samples were diluted and incubated, followed by detection with digoxygenin (dig-) conjugated 3E7 and HRP-conjugated sheep Fab anti-Dig antibodies (from Roche Applied Science, Mannheim, Germany), respectively. Enzyme activity was developed using ABTS.
Th e optical density was measured using a microplate reader, and results were calculated on basis of a calibration line using pooled normal human serum with a known concentration of MBL.
To estimate the eff ect of the surgical procedure on the MBL serum concentration, the ratio of the post- and preoperative MBL serum level was calculated.
Statistics
Categorical characteristics among MBL groups were compared using cross-tables with calculation of p-values. Continuous variables were analyzed using Student t-test or, when
appropriate Mann -Whitney test was used. Th e diff erences in MBL levels before and aft er
surgery were analyzed by the Wilcoxon’s signed rank test. Spearman Rank correlation
coeffi cients were used for correlation. A logistic regression model was used to evaluate the
eff ect of per-operative factors on the postoperative/preoperative MBL ratio. P-values <0.05 were considered as statistically signifi cant. All analyses were performed in SPSS (SPSS Inc, Chicago, IL, USA).
RESULTS
Patients Characteristics
Of the 474 patients included in the trial, serum for MBL determination was available from 400 patients prior to surgery. From these 400 patients, postoperative MBL levels could be
assessed in 330 patients. Th e main characteristics of the patient population are presented in
Table 1.
Serum MBL Levels
Pre-operative serum levels of MBL were highly variable as expected in the human population (range 15-3709 ng/ml; Table 1). In Figure 1A the post-/ preoperative MBL ratios are shown.
Th e ratio of post-operative and pre-operative MBL concentration was negatively correlated
with the pre-operative MBL concentration (R = -0.73, p<0.0001). Patients with high serum MBL levels before surgery (>400 ng/ml) showed an average decrease of 52 ± 12% of the serum MBL level aft er the procedure (range ratio 0.19-0.93; p < 0.0001, Figure 1B). In contrast, patients with low serum MBL levels before operation (< 400 ng/ml) generally showed an increase of serum MBL levels, as represented by a ratio signifi cantly higher than 1 (p<0.0001, Figure 1B).
Table 1 | Patient Characteristics Total Randomization* Buff y-coat depleted red blood cells Leuko- reduced red blood cells P value Patient number 400 194 206 Age (mean ±SD) 65.8 ± 13.9 66.5 ± 12.5 65.1 ± 15 0.36 Type of surgery (N (%)) Valve Valve + CABG 269 (67.3%) 131 (32.8%) 133 (68.6%) 61 (31.4%) 136 (66.0%) 70 (34.0%) 0.60 Female (N (%)) 178 (44.5%) 85 (43.8%) 93 (45.1%) 0.84 Parsonnet score 13.8 ± 8.2 13.8 ± 8.6 13.8 ± 7.8 0.98 Blood transfusions during surgery†
Red blood cells (mean units per patient ±SD) FFP (mean units per patient ±SD)
3.2 ± 3.6 3.2 ± 4.1 3.4 ± 4.7 3.6 ± 5.4 3.0 ± 2.1 2.8 ± 1.7 0.23 0.22 Patients without FFP transfusions (N (%))
Patients without red blood cell transf. (N (%))
224 (56.0%) 114 (28.5%) 103 (53.1%) 54 (27.8%) 121 (58.7%) 60 (29.1%) 0.27 0.82 Cardiopulmonary bypass time (minutes; mean ±SD) 140 ± 62 145 ± 66 136 ± 57 0.13 Aortic cross clamping time (minutes; mean ±SD) 96± 46 96 ± 47 96 ± 44 0.98 Serum total protein (gram/l) (mean ±SD)
Pre-operative Post-operative 74.5 ± 4.3 49.9 ± 6.2 74.3 ± 4.0 49.6 ± 6.6 75 ± 4.6 50.2 ± 5.8 0.55 0.50 Infections (N (%)) 113 (28.3%) 66 (34.0%) 47 (22.8%) 0.02 MODS (N (%)) 89 (22.3%) 45 (23.2%) 44 (21.4%) 0.72 Sepsis (N (%)) 17 (4.3%) 10 (5.2%) 7 (3.4%) 0.46 Hospital mortality (N (%)) 34 (8.5%) 23 (11.9%) 11 (5.3%) 0.04 Mean MBL (ng/ml) (mean ±SD) Pre-operative (N=400) Post-operative (N=330) 837 ± 796 453 ± 350 820 ± 802 449 ± 356 854 ± 792 457 ± 346 0.68 0.82 Median MBL (range) Pre-operative Postoperative 571 (15-3709) 339 (35-1934) 534 (21-3619) 329 (35-1934) Pre-operative MBL <80 (ng/ml) (N (%)) Pre-operative MBL <400 (ng/ml) (N (%)) 38 (9.5%) 164 (41.0%) 21 (10.8%) 81 (41.7%) 17 (8.2%) 83 (40.3%) 0.40 0.84
* Patients were randomized to receive either buff y-coat-depleted red blood cells or leukocyte depleted red blood cells, when needed. † Units per patients transfused during surgery. Only patients who received red blood cells and FFP transfusions are included for calculation of mean and SD.
Figure 1 | Th e eff ect of plasma transfusion on levels of serum MBL in patients who undergo
cardiac surgery. A. Ratios of post/pre-surgery MBL levels were plotted against pre-operative serum MBL
levels in all patients (N=330). Th e dashed line indicates a serum MBL level of 400 ng/ml. B. Mean ratios post/ pre-surgery MBL levels ± SD were calculated for patients above and below 400 ng/ml who did or did not receive FFP during the operation, as indicated. ** indicates statistical signifi cance. Ratios in all groups were signifi cantly diff erent from 1 (P<0.0001; Wilcoxon signed rank test). Th e diff erence between patients with serum MBL level below 400 ng/ml who did or did not receive plasma was tested with the Mann Whitney test (P=0.0039).
Ratios of MBL levels were not diff erent between recipients of leukoreduced or buff y-coat depleted red blood cells. In both groups the MBL level decreased or increased with the same ratios determined by the preoperative MBL level (not shown). In contrast, FFP transfusions were associated with an additional increase of post-operative serum MBL in patients with a pre-operative MBL level below 400 ng/ml (median post/pre MBL ratio 1.25 versus 1.61,
p=0.0039; Figure1B), whereas FFP transfusion did not show an eff ect in patients with MBL levels above 400 ng/ml.
Following surgery, all patients displayed a dilution eff ect, as shown by a decreased serum total protein concentration (average ratio total protein post/pre-surgery =0.67 ± 0.08; range 0.46-0.86). However, this dilution factor was not correlated to post/pre-surgery MBL ratios, also not when patients with pre-MBL levels above 400 ng/ml were analysed only (R<0.1).
In a logistic regression model we further evaluated the contribution of various per-
operative factors on the post/pre-surgery MBL ratio. Th e type of surgery (p=0.23), duration
of cardiopulmonary bypass (p=0.19), duration of aortic crossclamping (p=0.25), number of red blood cell transfusions (p=0.19), randomization arm (p=0.74) and the decrease in total protein concentrations (p=0.43) did not show a signifi cant eff ect on the post/preoperative MBL ratio.
MBL Serum Levels and Complications of Cardiac Surgery
Following cardiac surgery 113 of 400 patients developed infections (28.3%), 89 patients developed MODS (22.3%), 17 patients had sepsis (4.3%) and 34 patients died (8.5%; Table 1). Infections and mortality were signifi cantly higher in patients receiving leukocyte- containing transfusions than in patients randomized for leuko-reduced red blood cells (p=0.02 and p=0.04, respectively). Patients who developed MODS had a signifi cantly longer duration of cardiopulmonary bypass (median 148 versus 124 minutes; p=0.0002) and aortic crossclamping (median 98 versus 90 minutes; p=0.008).
Since data as presented above show that the surgical procedure and plasma transfusions have an eff ect on serum MBL levels, the relation between serum MBL and complications of cardiac surgery were further evaluated based on pre-operative MBL levels in patients who did not receive plasma. Using an MBL level of 400 ng/ml as a cut-off value, no signifi cant association could be identifi ed between higher or lower pre-operative MBL levels and infections, sepsis, MODS or mortality in the 224 patients who did not receive FFP transfusions (Table 2). However, patients with an MBL level below 80 ng/ml, indicating overt
MBL defi ciency, did not develop MODS (Figure 2 and Table 2; p=0.016). Th e incidence of
infections, sepsis and mortality was not signifi cantly diff erent in patients with pre-operative MBL serum levels below or above 80 ng/ml.
In total 176 patients received FFP transfusions. In patients with pre-operative MBL levels below 80 ng/ml, the FFP transfusions increased the risk for development of MODS until the same degree as for patients with preoperative MBL concentrations above 80 ng/ml
(Table 3; p=0.048). In patients with pre-operative serum MBL levels above 80 ng/ml or above 400 ng/ml, FFP transfusion did not show any eff ect on the development of MODS.
Table 2 | Complications in Relation to Pre-operative MBL Levels in Patients who did not Receive Plasma
MBL (ng/ml)* All ≤80 >80 p-value§ ≤400 >400 p-value§
N† 224 18 206 93 131
Infection (%)‡ 25.0 22.2 25.2 1.0 28.0 22.9 0.43
Sepsis (%)‡ 3.6 0 3.9 1.0 5.4 2.3 0.28
MODS (%)‡ 21.0 0 22.8 0.016 19.4 22.1 0.74
Mortality (%)‡ 6.3 0 6.8 0.61 7.5 5.3 0.58
*MBL level before surgery; †Number of patients in each category; ‡Percentage of patients who developed this complication; §P value (Fisher exact test; comparison between MBL-low and MBL-high patients); MODS = multiple-organ-dysfunction-syndrome
Figure 2 | Low serum MBL levels protect against MODS in patients who did not receive plasma during operation. Pre-operative serum MBL levels are shown in patients who did or did not develop MODS. Only patients who did not receive FFP were plotted.
Table 3 | Plasma Transfusion is a Risk Factor for MODS in Patients with Low Pre-operative MBL Levels
MBL level (ng/ml)* No plasma (N)† % MODS‡ Plasma (N)† % MODS‡ p-value§
All 224 21 176 24 0.55
≤80 18 0 20 25.0 0.048
>80 206 22.8 156 23.7 0.90
≤400 93 19.4 7 1 25.3 0.45
>400 131 22.1 105 22.9 1.00
* MBL level before surgery; †Number of patients who did or did not receive plasma during the procedure; ‡Percentage of patients who developed multi-organ dysfunction syndrome; §P value (Fisher exact test; comparison between patients who did or did not receive plasma)
CONCLUSIONS
In a randomised study in cardiac surgery patients we observed a lower incidence of postoperative infections and mortality associated with MODS aft er transfusion of fi ltered leukoreduced red blood cell transfusions compared with standard buff y-coat depleted
red blood cells [17]. Pre-operative and postoperative blood samples had been taken for
investigation of mechanisms, which could explain diff erences between both randomization arms, should these occur. Because in previous studies low MBL levels have been associated
with infections and high levels with ischemia/reperfusion injury [6-12], both important
complications of cardiac surgery, we hypothesized that leukocytes in red blood cell units may play a role by activation of the innate immune system via the lectin pathway.
Th e results of this study show that absence or presence of allogeneic leukocytes
in erythrocyte products had no eff ect on MBL concentrations aft er surgery. In both randomisation arms, patients with higher (>400 ng/ml) pre-operative MBL levels show post- surgical decrease by almost 50%. Although patients with preoperative MBL values below 400 ng/ml increased their postoperative MBL levels, especially if FFP had been administered during the procedure. If the preoperative MBL value was above 80 ng/ml, preoperative and postoperative MBL values were not associated with postoperative infections, MODS or mortality. However, none of the 18 patients with MBL defi ciency below 80 ng/ml who did not receive per-operative FFP developed MODS. In contrast, the incidence of MODS in 20 MBL-defi cient patients who received FFP was comparable with the incidence in patients with higher preoperative MBL concentrations.
Th is is, to the best of our knowledge, the fi rst study that reports MBL serum levels aft er cardiac surgery. In a previous study in a small number of 23 patients undergoing abdominal aneurysm aorta surgery, a postoperative decrease of approximately 40% of MBL was observed, whereas in a control group, undergoing biliary and pancreatic surgery, there was no eff ect on
MBL levels aft er the operation [23]. In patients undergoing abdominal surgery for respectable
esophagus carcinoma a clear increase of serum MBL was observed, with a slow kinetics
starting on day 5 postoperatively [24]. In gastric surgery patients, no eff ect on post-surgery
MBL levels was observed as assessed on day 3 [25]. Th ese variable results in patient cohorts
undergoing diff erent surgical interventions, in studies with relatively small cohort sizes, indicate interplay of diff erent factors modifying the MBL serum level. In our large patient population undergoing cardiac surgery, we observed an increase as well as a decrease of MBL
concentrations. Th e change in MBL concentration aft er surgery was signifi cantly associated
with the patient’s MBL status prior to surgery. Because the MBL level in the population shows a wide range, which is genetically based, this relationship is easily missed when small
groups of patients are evaluated. Th e measured decrease of 50% in the MBL concentration
was somewhat overestimated because of postoperative hemodilution, refl ected by a decrease of the total protein concentration. In reverse, due to this hemodilution, the increase of MBL aft er cardiac surgery is somewhat underestimated. When calculation of MBL post/pre ratios was corrected for this hemodilution eff ect, we confi rmed a median 24% decrease of serum MBL levels (P<0.0001, not shown) in patients with pre-operative MBL levels above 400 ng/ml, suggesting an independent cause for the decrease of serum MBL.
Th e reduction in serum MBL in patients with higher preoperative MBL levels is
most likely the result of consumption by activation of the lectin pathway during surgery, presumably mainly by interaction of MBL with ischemic and injured tissue. Accordingly, MBL consumption was previously observed in patients undergoing aorta aneurism operation
[23], a procedure also associated with extended ischemia, but not in patients undergoing
other surgical interventions [23-25]. Th e fi nding that this decrease was not compensated in
our cardiac surgery patients, neither by endogenous production nor by FFP transfusions, suggests a stronger MBL consumption in these patients than in patients with lower preoperative MBL levels. In patients with preoperative MBL concentrations below 400 ng/ ml we do not know whether (compensated) per-operative MBL consumption may occur, but aft er surgery endogenous production is increased and the level further increases aft er FFP
transfusions. Th e latter observation, i.e. MBL reconstitution of MBL-defi cient patients upon
FFP transfusion, has also been reported for patients undergoing aortic aneurysm repair [3].
level before transplantation, since levels of MBL above 400 ng/ml are strongly associated with the MBL wildtype genotype and normal lectin pathway function, whereas MBL from
low-MBL patients may show impaired interaction with ligand [26]. Further studies with
sampling during the various phases in cardiac surgery are needed revealing the kinetics of MBL during surgery to explain the mechanisms of this dichotic MBL behavior.
In the present study no association was observed between pre- or post-surgery MBL levels and postoperative infections aft er cardiac surgery, in contrast to studies performed in immunocompromised patients showing an association of MBL defi ciency with increased
risk of infections [6-9]. However, patients with an obvious MBL defi ciency before surgery
were signifi cantly protected against the development of MODS. Only in these patients, FFP transfusions, resulting into (partial) reconstitution of the MBL status, reversed this
protection against MODS. Th is is in agreement with observations in experimental renal
and myocardial ischemia/reperfusion injury and septicaemia in rats and mice, which suggest that low MBL ameliorates ischemia/reperfusion injury and sepsis-induced organ damage
[12,13,27,28]. Lower MBL levels have also been associated with less graft loss in kidney
transplantation in humans [22]. In the setting of aortic aneurism repair, one patient has been
presented who was MBL defi cient and who did not respond to the surgical procedure with production of soluble complement activation products and cytokines, in contrast to all MBL-
suffi cient patients who developed an evident infl ammatory reaction [3]. MBL may mediate
an infl ammatory response via activation of the complement cascade and probably also by
direct interaction with and activation of leukocytes [19]. However we observed a benefi cial
eff ects of transfusion with leukoreduced red blood cells in cardiac surgery [17], no eff ects of
allogeneic leukocytes could be observed in this study in postoperative MBL levels. Taken together, results now presented in our study support a role for MBL in the activation of the complement system and induction of a systemic infl ammatory response upon ischemia and reperfusion injury, which may evolve into multiple organ dysfunction. MBL defi ciency have probably a protective role in development of MODS, which disappears with transfusions of FFP and is not directly related with allogeneic leukocytes.
ACKNOWLEDGMENTS
Th e authors thank Dr. T. Fujita (Fukushima, Japan) for kindly providing reagents used in the
REFERENCES
1. Kollef MH, Wragge T, Pasque C. Determinants of mortality and multiorgan dysfunction in cardiac surgery patients requiring prolonged mechanical ventilation. Chest 1995; 107:1395-1401.
2. Wan S, LeClerc JL, Vincent JL. Infl ammatory response to cardiopulmonary bypass. Chest 1997; 112:676- 92.
3. Fiane AE, Videm V, Lingaas PS, Heggelund L, Nielsen EW, Geiran OR, Fung M, Mollnes TE. Mechanism of complement activation and its role in the infl ammatory response aft er thoracoabdominal aortic aneurysm repair. Circulation 2003; 108:849-56.
4. Neth O, Jack DL, Dodds AW, Holzel H, Klein NJ, Turner MW. Mannose-binding lectin binds to a range of clinically relevant microorganisms and promotes complement deposition. Infection and immunity 2000; 68:688-93.
5. Dahl M, Tybjaerg-Hansen A, Schnohr P, Nordestgaard BG. A population-based study of morbidity and mortality in MBL-defi ciency. J Exp Med 2004; 199:1391-9.
6. Mullighan CG, Heatley S, Doherty K, Szabo F, Grigg A, Hughes TP, Schwarer AP, Szer J, Tait BD, Bik To L, Bardy PG. Mannose-binding lectin gene polymorphisms are associated with major infection following allogeneic hemopoietic stem cell transplantation. Blood 2002; 10:3524-9.
7. Peterslund NA, Koch C, Jensenius JC, Th iel S. Association between defi ciency of mannose-binding lectin and severe infections aft er chemotherapy. Lancet 2001; 358:637-8.
8. Garred P, Madsen HO, Balslev U, Hofmann B, Pedersen C, Gerstoft J, Svejgaard A. Suspectibility to HIV infection and progression of AIDS in relation to variant alleles of mannose-binding lectin. Lancet 1997; 349:236-40.
9. Bouwman LH, Roos A, Terpstra OT, de Knijff P, van Hoek B, Verspaget HW, Berger SP, Daha MR, Frölich M, van der Slik AR, Doxiadis II, Roep BO, Schaapherder AF. Mannose binding lectin gene polymorphisms confer a major risk for severe infections aft er liver transplantation. Gastroenterology 2005; 129:408-14.
10. Fidler KJ, Wilson P, Davies JC, Turner MW, Peters MJ, Klein NJ. Increased incidence and severity of the systemic infl ammatory response syndrome in patients defi cient in mannose-binding lectin. Intensive Care Med 2004;30:1438-45.
11. Garred PJ, Strom JJ, Quist L, Taaning E, Madsen HO. Association mannose-binding lectin polymorphisms with sepsis and fatal outcome in patients with systemic infl ammatory response syndrome. J Infect Dis 2003; 188:1394-1403.
12. Jordan JE, Montalto MC, Stahl GL. Inhibition of mannose-binding lectin reduces postischemic myocardial reperfusion injury. Circulation 2001;104:1413-8.
13. Walsh MC, Bourcier T, Takahashi K, Busche MN, Rother RP, Solomon SD, Ezekowitz RA, Stahl GL. Mannose-binding lectin is a regulator of infl ammation that accompanies myocardial ischemia and